Empowering Autism Therapy Centers with a Retrofit and Computer Vision Based Solution for Behavioral Monitoring and Actionable Insights

Autism is a neurological disorder where social and motor skills are compromised. However, therapies could help to a great extent to improve these skills if performed and monitored properly. In the context of autism therapy centers, the underutilization of existing CCTV resources represents a missed opportunity for enhancing therapeutic practices. This paper introduces a technology-driven approach to address this gap by integrating advanced machine learning with existing infrastructure. Our proposed system leverages multi-variate LSTM architecture, validated on the SSDB dataset with an accuracy of 86%, to augment behavioral analysis in an effective and scalable manner. Through our research, we aim to showcase how technological integration can significantly improve therapeutic outcomes for individuals with autism, setting the stage for future advancements in technology-assisted therapies and broader applications in supporting neurodiversity.